Numerous natural and synthetic proteins are used in diagnostic and therapeutic applications; many others are in development or in clinical trials. Current methods of protein production include isolation from natural sources and recombinant production in bacterial and mammalian cells. Because of the complexity and high cost of these methods of protein production, however, efforts are underway to develop alternatives. For example, methods for producing exogenous proteins in the milk of pigs, sheep, goats, and cows have been reported. These approaches have certain limitations, including long generation times between founder and production transgenic herds, extensive husbandry and veterinary costs, and variable levels of expression because of position effects at the site of the transgene insertion in the genome. Proteins are also being produced using milling and malting processes from barley and rye. However, plant post-translational modifications differ from vertebrate post-translational modifications, which often has a critical effect on the function of the exogenous proteins such as pharmaceutical proteins.
Like tissue culture and mammary gland bioreactors, the avian oviduct can also potentially serve as a bioreactor. Successful methods of modifying avian genetic material such that high levels of exogenous proteins are secreted in the oviduct and packaged into eggs would allow inexpensive production of large amounts of protein. Several advantages of such an approach would be: a) short generation times (24 weeks) and rapid establishment of transgenic flocks via artificial insemination; b) readily scaled production by increasing flock sizes to meet production needs; c) post-translational modification of expressed proteins; 4) automated feeding and egg collection; d) naturally sterile egg-whites; and e) reduced processing costs due to the high concentration of protein in the egg white.
The avian reproductive system, including that of the chicken, is well described. The egg of the hen consists of several layers which are secreted upon the yolk during its passage through the oviduct. The production of an egg begins with formation of the large yolk in the ovary of the hen. The unfertilized oocyte is then positioned on top of the yolk sac. Upon ovulation or release of the yolk from the ovary, the oocyte passes into the infundibulum of the oviduct where it is fertilized if sperm are present. It then moves into the magnum of the oviduct which is lined with tubular gland cells. These cells secrete the egg-white proteins, including ovalbumin, lysozyme, ovomucoid, conalbumin, and ovomucin, into the lumen of the magnum where they are deposited onto the avian embryo and yolk.
The ovalbumin gene encodes a 45 kD protein that is specifically expressed in the tubular gland cells of the magnum of the oviduct (Beato Cell 56:335-344 (1989)). Ovalbumin is the most abundant egg white protein, comprising over 50 percent of the total protein produced by the tubular gland cells, or about 4 grams of protein per large Grade A egg (Gilbert, “Egg albumen and its formation” in Physiology and Biochemistry of the Domestic Fowl, Bell and Freeman, eds., Academic Press, London, N.Y., pp. 1291-1329). The ovalbumin gene and over 20 kb of each flanking region have been cloned and analyzed (Lai et al., Proc. Natl. Acad. Sci. USA 75:2205-2209 (1978); Gannon et al., Nature 278:428-424 (1979); Roop et al., Cell 19:63-68 (1980); and Royal et al., Nature 279:125-132 (1975)).
Much attention has been paid to the regulation of the ovalbumin gene. The gene responds to steroid hormones such as estrogen, glucocorticoids, and progesterone, which induce the accumulation of about 70,000 ovalbumin mRNA transcripts per tubular gland cell in immature chicks and 100,000 ovalbumin mRNA transcripts per tubular gland cell in the mature laying hen (Palmiter, J. Biol. Chem. 248:8260-8270 (1973); Palmiter, Cell 4:189-197 (1975)). DNAse hypersensitivity analysis and promoter-reporter gene assays in transfected tubular gland cells defined a 7.4 kb region as containing sequences required for ovalbumin gene expression. This 5′ flanking region contains four DNAse I-hypersensitive sites centered at −0.25, −0.8, −3.2, and −6.0 kb from the transcription start site. These sites are called HS-I, -II, -III, and -IV, respectively. These regions reflect alterations in the chromatin structure and are specifically correlated with ovalbumin gene expression in oviduct cells (Kaye et al., EMBO 3:1137-1144 (1984)). Hypersensitivity of HS-II and -III are estrogen-induced, supporting a role for these regions in hormone-induction of ovalbumin gene expression.
HS-I and HS-II are both required for steroid induction of ovalbumin gene transcription, and a 1.4 kb portion of the 5′ region that includes these elements is sufficient to drive steroid-dependent ovalbumin expression in explanted tubular gland cells (Sanders and McKnight, Biochemistry 27: 6550-6557 (1988)). HS-I is termed the negative-response element (“NRE”) because it contains several negative regulatory elements which repress ovalbumin expression in the absence of hormones (Haekers et al., Mol. Endo. 9:1113-1126 (1995)). Protein factors bind these elements, including some factors only found in oviduct nuclei suggesting a role in tissue-specific expression. HS-II is termed the steroid-dependent response element (“SDRE”) because it is required to promote steroid induction of transcription. It binds a protein or protein complex known as Chirp-I. Chirp-I is induced by estrogen and turns over rapidly in the presence of cyclohexamide (Dean et al., Mol. Cell. Biol. 16:2015-2024 (1996)). Experiments using an explanted tubular gland cell culture system defined an additional set of factors that bind SDRE in a steroid-dependent manner, including an NFκB-like factor (Nordstrom et al., J. Biol. Chem. 268:13193-13202 (1993); Schweers and Sanders, J. Biol. Chem. 266: 10490-10497 (1991)).
Less is known about the function of HS-III and -IV. HS-III contains a functional estrogen response element, and confers estrogen inducibility to either the ovalbumin proximal promoter or a heterologous promoter when co-transfected into HeLa cells with an estrogen receptor cDNA. These data imply that HS-III may play a functional role in the overall regulation of the ovalbumin gene. Little is known about the function of HS-IV, except that it does not contain a functional estrogen-response element (Kato et al., Cell 68: 731-742 (1992)).
There has been much interest in modifying eukaryotic genomes by introducing foreign genetic material and/or by disrupting specific genes. Certain eukaryotic cells may prove to be superior hosts for the production of exogenous eukaryotic proteins. The introduction of genes encoding certain proteins also allows for the creation of new phenotypes which could have increased economic value. In addition, some genetically-caused disease states may be cured by the introduction of a foreign gene that allows the genetically defective cells to express the protein that they can otherwise not produce. Finally, modification of animal genomes by insertion or removal of genetic material permits basic studies of gene function, and ultimately may permit the introduction of genes that could be used to cure disease states, or result in improved animal phenotypes.
Transgenesis has been accomplished in mammals by several different methods. First, in mammals including the mouse, pig, goat, sheep and cow, a transgene is microinjected into the pronucleus of a fertilized egg, which is then placed in the uterus of a foster mother where it gives rise to a founder animal carrying the transgene in its germline. The transgene is engineered to carry a promoter with specific regulatory sequences directing the expression of the foreign protein to a particular cell type. Since the transgene inserts randomly into the genome, position effects at the site of the transgene's insertion into the genome may variably cause decreased levels of transgene expression. This approach also requires characterization of the promoter such that sequences necessary to direct expression of the transgene in the desired cell type are defined and included in the transgene vector (Hogan et al. Manipulating the Mouse Embryo, Cold Spring Harbor Laboratory, NY (1988)).
A second method for effecting animal transgenesis is targeted gene disruption, in which a targeting vector containing sequences of the target gene flanking a selectable marker gene is introduced into embryonic stem (“ES”) cells. By homologous recombination, the targeting vector replaces the target gene sequences at the chromosomal locus or inserts into interior sequences preventing expression of the target gene product. Clones of ES cells having the appropriately disrupted gene are selected and then injected into early stage blastocysts generating chimeric founder animals, some of which have the transgene in the germ line. In the case where the transgene deletes the target locus, it replaces the target locus with foreign DNA borne in the transgene vector, which consists of DNA encoding a selectable marker useful for detecting transfected ES cells in culture and may additionally contain DNA sequences encoding a foreign protein which is then inserted in place of the deleted gene such that the target gene promoter drives expression of the foreign gene (U.S. Pat. Nos. 5,464,764 and 5,487,992 (M. P. Capecchi and K. R. Thomas)). This approach suffers from the limitation that ES cells are unavailable in many mammals, including goats, cows, sheep and pigs. Furthermore, this method is not useful when the deleted gene is required for survival or proper development of the organism or cell type.
Recent developments in avian transgenesis have allowed the modification of avian genomes. Germ-line transgenic chickens may be produced by injecting replication-defective retrovirus into the subgerminal cavity of chick blastoderms in freshly laid eggs (U.S. Pat. No. 5,162,215; Bosselman et al., Science 243:533-534 (1989); Thoraval et al., Transgenic Research 4:369-36 (1995)). The retroviral nucleic acid carrying a foreign gene randomly inserts into a chromosome of the embryonic cells, generating transgenic animals, some of which have the transgene in their germ line. Use of insulator elements inserted at the 5′ or 3′ region of the fused gene construct to overcome position effects at the site of insertion has been described (Chim et al., Cell 74:504-514 (1993)).
In another approach, a transgene has been microinjected into the germinal disc of a fertilized egg to produce a stable transgenic founder avian that may pass the gene to the F1 generation (Love et al., Bio/Technology 12:60-63 (1994)). However, this method has several disadvantages. Hens must be sacrificed in order to collect the fertilized egg, the fraction of transgenic founders is low, and injected eggs require labor intensive in vitro culture in surrogate shells.
In another approach, blastodermal cells containing presumptive primordial germ cells (“PGCs”) are excised from donor eggs, transfected with a transgene and introduced into the subgerminal cavity of recipient embryos. The transfected donor cells are incorporated into the recipient embryos generating transgenic embryos, some of which are expected to have the transgene in the germ line. The transgene inserts in random chromosomal sites by nonhomologous recombination. However, no transgenic founder avians have yet been generated by this method.
Lui, Poult. Sci. 68:999-1010 (1995), used a targeting vector containing flanking DNA sequences of the vitellogenin gene to delete part of the resident gene in chicken blastodermal cells in culture. However, it has not been demonstrated that these cells can contribute to the germ line and thus produce a transgenic embryo. In addition, this method is not useful when the deleted gene is required for survival or proper development of the organism or cell type.
Thus, it can be seen that there is a need for a method of introducing foreign DNA, operably linked to a suitable promoter, into the avian genome such that efficient expression of an exogenous gene can be achieved. Furthermore, there exists a need to create germ-line modified transgenic avians which express exogenous genes in their oviducts and secrete the expressed exogenous proteins into their eggs.
When interferon was discovered in 1957, it was hailed as a significant antiviral agent. In the late 1970s, interferon became associated with recombinant gene technology. Today, interferon is a symbol of the complexity of the biological processes of cancer and the value of endurance and persistence in tackling this complexity.
The abnormal genes that cause cancer comprise at least three types: firstly, there are the oncogenes, which, when altered, encourage the abnormal growth and division that characterize cancer. Secondly, there are the tumor suppressor genes, which, when altered, fail to control this abnormal growth and division. Thirdly, there are the DNA repair genes, which, when altered, fail to repair mutations that can lead to cancer. Researchers speculate that there are about 30 to 40 tumor suppressor genes in the body, each of which produces a protein. These proteins may be controlled by “master” tumor suppressor proteins such as Rb (for retinoblastoma, with which it was first associated) and p53 (associated with many different tumors). Evidence from the laboratory suggests that returning just one of these tumor suppressor genes to its normal function can appreciably reduce the aggressiveness of the malignancy.
Scientists became intrigued by interferon when it was discovered that interferon can inhibit cell growth. Further, interferon was found to have certain positive effects on the immune system. It is now considered analogous to a tumor suppressor protein: it inhibits the growth of cells, particularly malignant cells; it blocks the effects of many oncogenes and growth factors; and unlike other biological agents, it inhibits cell motility which is critical to the process of metastasis.
Intercellular communication is dependent on the proper functioning of all the structural components of the tissue through which the messages are conveyed: the matrix, the cell membrane, the cytoskeleton, and the cell itself. In cancer, the communication network between cells is disrupted. If the cytoskeleton is disrupted, the messages don't get through to the nucleus and the nucleus begins to function abnormally. Since the nucleus is the site where the oncogenes or tumor suppressor genes get switched on or off, this abnormal functioning can lead to malignancy. When this happens, the cells start growing irregularly and do not differentiate. They may also start to move and disrupt other cells. It is believed that interferon, probably in concert with other extracellular and cellular substances, restores the balance and homeostasis, making sure the messages get through properly. Interferon stops growth, stops motility, and enhances the ability of the cell, through adhesion molecules, to respond to its environment. It also corrects defects and injuries in the cytoskeleton. Interferon has been found to block angiogenesis, the initial step in the formation of new blood vessels that is essential to the growth of malignancies. Moreover, it blocks fibrosis, a response to injury that stimulates many different kinds of cells and promotes cell growth (Kathryn L. Hale, Oncolog, Interferon: The Evolution of a Biological Therapy, Taking a New Look at Cytokine Biology).
Interferon is produced by animal cells when they are invaded by viruses and is released into the bloodstream or intercellular fluid to induce healthy cells to manufacture an enzyme that counters the infection. For many years the supply of human interferon for research was limited by costly extraction techniques. In 1980, however, the protein became available in greater quantities through genetic engineering (i.e., recombinant forms of the protein). Scientists also determined that the body makes three distinct types of interferon, referred to as α-(alpha), β-(beta), and γ-(gamma) interferon. Interferons were first thought to be highly species-specific, but it is now known that individual interferons may have different ranges of activity in other species. Alpha interferon (α-IFN) has been approved for therapeutic use against hairy-cell leukemia and hepatitis C. α-IFN has also been found effective against chronic hepatitis B, a major cause of liver cancer and cirrhosis, as well as for treatment of genital warts and some rarer cancers of blood and bone marrow. Nasal sprays containing α-IFN provide some protection against colds caused by rhinoviruses. Human α-IFN belongs to a family of extra-cellular signaling proteins with antiviral, antiproliferating and immunomodulatory activities. IFN-α proteins are encoded by a multigene family which includes 13 genes clustered on the human chromosome 9. Most of the IFN-α genes are expressed at the mRNA level in leukocytes induced by Sendai virus. Further, it has been shown that at least nine different sub-types are also produced at the protein level. The biological significance of the expression of several similar IFN-α proteins is not known, however, it is believed that they have quantitatively distinct patterns of antiviral, growth inhibitory and killer-cell-stimulatory activities. Currently, two IFN-α variants, IFN-α 2a and IFN-α 2b, are mass produced in Escherichia coli by recombinant technology and marketed as drugs.
Unlike natural IFN-α, these recombinant IFN-α products have been shown to be immunogenic in some patients, which could be due to unnatural forms of IFN-α proteins. Thus, for the development of IFN-α drugs it is necessary to not only identify the IFN-α subtypes and variants expressed in normal human leukocytes, but also to characterize their possible post-translational modifications (Nyman et al. (1998) Eur. J. Biochem. 253:485-493).
Nyman et al. (supra) studied the glycosylation of natural human IFN-α. They found that two out of nine of the subtypes produced by leukocytes after a Sendai-virus induction were found to be glycosylated, namely IFN-α 14c and IFN-α 2b, which is consistent with earlier studies. IFN-α 14 is the only IFN-α subtype with potential N-glycosylation sites, Asn2 and Asn72, but only Asn72 is actually glycosylated. IFN-α 2 is O-glycosylated at Threonine 106 (Thr106). Interestingly, no other IFN-α subtype contains Thr at this position. In this study, Nyman et al. liberated and isolated the oligosaccharide chains and analyzed their structures by mass spectrometry and specific glycosidase digestions. Both IFN-α 2b and IFN-α 14c resolved into three peaks in reversed-phase high performance liquid chromatography (RP-HPLC). Electrospray ionization mass spectrometry (ESI-MS) analysis of IFN-α 2b fractions from RP-HPLC revealed differences in their molecular masses, suggesting that these represent different glycoforms. This was confirmed by masspectrometric analysis of the liberated O-glycans of each fraction. IFN-α 2b was estimated to contain about 20% of the core type-2 pentasaccharide, and about 50% of disialylated and 30% of monosialylated core type-1 glycans. Nyman et al.'s data agrees with previous partial characterization of IFN-α 2b glycosylation (Adolf et al. (1991) Biochem. J. 276:511-518). The role of glycosylation in IFN-α 14c and IFN-α 2b is not clearly established. According to Nyman et al. (supra), the carbohydrate chains are not essential for the biological activity, but glycosylation may have an effect on the pharmacokinetics and stability of the proteins.
There are at least 15 functional genes in the human genome that code for proteins of the IFN-α family. The amino acid sequence similarities are generally in the region of about 90%, thus, these molecules are closely related in structure. IFN-α proteins contain 166 amino acids (with the exception of IFN-α 2, which has 165 amino acids) and characteristically contain four conserved cysteine residues which form two disulfide bridges. IFN-α species are slightly acidic in character and lack a recognition site for asparagine-linked glycosylation (with the exception of IFN-α 14 which does contain a recognition site for asparagine-linked glycosylation). Three variants of IFN-α 2, differing in their amino acids at positions 23 and 34, are known: IFN-α 2a (Lys-23, His-34); IFN-α 2b (Arg-23, His-34); and IFN-α 2c (Arg-23, Arg-34). It is believed that IFN-α 2a and IFN-α 2c are allelic variants of IFN-α 2b. See, Gewert et al (1993) J. Interferon Res. vol 13, p 227-231. The minor differences in amino acid content of the IFN-α 2 species is not expected to effect glycosylation of the interferons. That is glycosyation patterns are expected to be essentially the same for each of IFN-α 2a, 2b and 2c. Two other human IFN species, namely IFN-ω 1 and IFN-β are N-glycosylated and are more distantly related to IFN-α. IFN-α, -β and -ω, collectively referred to as class I IFNs, bind to the same high affinity cell membrane receptor (Adolf et al. (1991) Biochem. J. 276:511-518).
Adolf et al. (supra) used the specificity of a monoclonal antibody for the isolation of natural IFN-α 2 from human leukocyte IFN. They obtained a 95% pure protein through immunoaffinity chromatography which confirmed the expected antiviral activity of IFN-α 2. Analysis of natural IFN-α 2 by reverse-phase HPLC, showed that the natural protein can be resolved into two components, both more hydrophilic than E. coli-derived IFN-α 2. SDS/PAGE revealed that the protein is also heterogeneous in molecular mass, resulting in three bands, all of them with lower electrophoretic mobility than the equivalent E. coli-derived protein.
Adolf et al. (supra) also speculated that natural IFN-α 2 carries O-linked carbohydrate residues. Their hypothesis was confirmed by cleavage of the putative peptide-carbohydrate bond with alkali; the resulting protein was homogeneous and showed the same molecular mass as the recombinant protein. Further comparison of natural and recombinant proteins after proteolytic cleavage, followed by separation and analysis of the resulting fragments, allowed them to define a candidate glycopeptide. Sequence analysis of this peptide identified Thr-106 as the O-glycosylation site. A comparison of the amino acid sequences of all published IFN-α 2 species revealed that this threonine residue is unique to IFN-α 2. Glycine, isoleucine or glutamic acid are present at the corresponding position (107) in all other proteins.
Preparations of IFN-α 2 produced in E. coli are devoid of O-glycosylation and have been registered as drugs in many countries. However, the immunogenicity of therapeutically applied E. coli-derived IFN-α 2 might be affected by the lack of glycosylation. Studies have shown that four out of sixteen patients receiving recombinant human granulocyte-macrophage colony-stimulating factor produced in yeast developed antibodies to this protein. Interestingly, these antibodies were found to react with epitopes that in the endogenous granulocyte-macrophage colony-stimulating factor are protected by O-linked glycosylation, but which are exposed in the recombinant factor (Adolf et al., supra).
Similarly, induction of antibodies to recombinant E. coli-derived IFN-α 2 after prolonged treatment of patients has been described and it has been speculated that natural IFN-α 2 may be less immunogenic than the recombinant IFN-α 2 proteins (Galton et al. (1989) Lancet 2:572-573).
What is needed are improved methods of producing therapeutic or pharmaceutical proteins such as antibodies and cytokines including interferon, G-CSF and erythropoietin.